Fake news no GANs Do Deep Generative Models
![*Fake news, no GANs Do Deep Generative Models* Know What They Don't Know? Eric *Fake news, no GANs Do Deep Generative Models* Know What They Don't Know? Eric](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-1.jpg)
![TL; DR TL; DR](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-2.jpg)
![*in some interesting cases TL; DR Normalizing flows, VAEs, Pixel. CNNs aren’t reliable enough *in some interesting cases TL; DR Normalizing flows, VAEs, Pixel. CNNs aren’t reliable enough](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-3.jpg)
![Outline • Paper introduction • Some notes • How normalizing flows work? • Paper Outline • Paper introduction • Some notes • How normalizing flows work? • Paper](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-4.jpg)
![Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-5.jpg)
![Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-6.jpg)
![Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-7.jpg)
![Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-8.jpg)
![Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-9.jpg)
![Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows. Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-10.jpg)
![Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows. Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-11.jpg)
![Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows. Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-12.jpg)
![How normalizing flows work? How normalizing flows work?](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-13.jpg)
![How normalizing flows work? • *Illustration stolen from here: https: //www. youtube. com/watch? v=P How normalizing flows work? • *Illustration stolen from here: https: //www. youtube. com/watch? v=P](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-14.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-15.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-16.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-17.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-18.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-19.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-20.jpg)
![How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf): How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf):](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-21.jpg)
![How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf): How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf):](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-22.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-23.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-24.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-25.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-26.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-27.jpg)
![How normalizing flows work? • How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-28.jpg)
![Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-29.jpg)
![Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-30.jpg)
![Paper findings • Fashion. MNIST vs. MNIST Paper findings • Fashion. MNIST vs. MNIST](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-31.jpg)
![Paper findings • Fashion. MNIST vs. MNIST Paper findings • Fashion. MNIST vs. MNIST](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-32.jpg)
![Paper findings • CIFAR-10 vs. SVHN Paper findings • CIFAR-10 vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-33.jpg)
![Paper findings • CIFAR-10 vs. SVHN Paper findings • CIFAR-10 vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-34.jpg)
![Paper findings • Celeb. A vs. SVHN Paper findings • Celeb. A vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-35.jpg)
![Paper findings • Celeb. A vs. SVHN Paper findings • Celeb. A vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-36.jpg)
![Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-37.jpg)
![Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-38.jpg)
![Paper findings • Other model types Paper findings • Other model types](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-39.jpg)
![Paper findings • The observations presented were the main contributions of the paper, grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-40.jpg)
![Paper findings • The observations presented were the main contributions of the paper, grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-41.jpg)
![Paper findings • The observations presented were the main contributions of the paper, grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-42.jpg)
![Paper findings • Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-43.jpg)
![Paper findings • Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-44.jpg)
![Paper findings • Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-45.jpg)
![Paper findings • Then hypothesize that reducing the variance of the data artificially will Paper findings • Then hypothesize that reducing the variance of the data artificially will](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-46.jpg)
![Conclusions • Cause to pause when using generative models in anomaly detection • Second Conclusions • Cause to pause when using generative models in anomaly detection • Second](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-47.jpg)
![Discussion • How valid/applicable is their analysis? • How come samples do not look Discussion • How valid/applicable is their analysis? • How come samples do not look](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-48.jpg)
- Slides: 48
![Fake news no GANs Do Deep Generative Models Know What They Dont Know Eric *Fake news, no GANs Do Deep Generative Models* Know What They Don't Know? Eric](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-1.jpg)
*Fake news, no GANs Do Deep Generative Models* Know What They Don't Know? Eric Nalisnick, Akihiro Matsukawa, Yee Whye Teh, Dilan Gorur, Balaji Lakshminarayanan (Deep. Mind) ICLR 2019 Presented by: Julius Hietala
![TL DR TL; DR](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-2.jpg)
TL; DR
![in some interesting cases TL DR Normalizing flows VAEs Pixel CNNs arent reliable enough *in some interesting cases TL; DR Normalizing flows, VAEs, Pixel. CNNs aren’t reliable enough](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-3.jpg)
*in some interesting cases TL; DR Normalizing flows, VAEs, Pixel. CNNs aren’t reliable enough to detect out of distribution data*
![Outline Paper introduction Some notes How normalizing flows work Paper Outline • Paper introduction • Some notes • How normalizing flows work? • Paper](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-4.jpg)
Outline • Paper introduction • Some notes • How normalizing flows work? • Paper experiments • Paper findings • Conclusions • Discussion
![Paper introduction Density estimationdetermination is used in many applications anomaly detection transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-5.jpg)
Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning etc. )
![Paper introduction Density estimationdetermination is used in many applications anomaly detection transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-6.jpg)
Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning etc. ) • These applications have spawned interest towards deep generative models
![Paper introduction Density estimationdetermination is used in many applications anomaly detection transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-7.jpg)
Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning etc. ) • These applications have spawned interest towards deep generative models • Currently popular choices are VAEs, GANs, auto regressive models, and invertible latent variable models
![Paper introduction Density estimationdetermination is used in many applications anomaly detection transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-8.jpg)
Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning etc. ) • These applications have spawned interest towards deep generative models • Currently popular choices are VAEs, GANs, auto regressive models, and invertible latent variable models • The latter two are interesting due to the fact that they allow for exact likelihood calculation
![Paper introduction Density estimationdetermination is used in many applications anomaly detection transfer learning Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-9.jpg)
Paper introduction • Density estimation/determination is used in many applications (anomaly detection, transfer learning etc. ) • These applications have spawned interest towards deep generative models • Currently popular choices are VAEs, GANs, auto regressive models, and invertible latent variable models • The latter two are interesting due to the fact that they allow for exact likelihood calculation • Main question of the paper: can these models be used for anomaly detection?
![Some notes The authors report results for VAEs Pixel CNNs and normalizing flows Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-10.jpg)
Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.
![Some notes The authors report results for VAEs Pixel CNNs and normalizing flows Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-11.jpg)
Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows. • Only normalizing flows are discussed and studied in depth
![Some notes The authors report results for VAEs Pixel CNNs and normalizing flows Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows.](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-12.jpg)
Some notes • The authors report results for VAEs, Pixel. CNNs, and normalizing flows. • Only normalizing flows are discussed and studied in depth • Is their analysis applicable to all the different types of models?
![How normalizing flows work How normalizing flows work?](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-13.jpg)
How normalizing flows work?
![How normalizing flows work Illustration stolen from here https www youtube comwatch vP How normalizing flows work? • *Illustration stolen from here: https: //www. youtube. com/watch? v=P](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-14.jpg)
How normalizing flows work? • *Illustration stolen from here: https: //www. youtube. com/watch? v=P 4 Ta-TZPVi 0
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-15.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-16.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-17.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-18.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-19.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-20.jpg)
How normalizing flows work? •
![How normalizing flows work Example from Real NVP https arxiv orgpdf1605 08803 pdf How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf):](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-21.jpg)
How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf): *s and t are NN()
![How normalizing flows work Example from Real NVP https arxiv orgpdf1605 08803 pdf How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf):](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-22.jpg)
How normalizing flows work? • Example from Real. NVP (https: //arxiv. org/pdf/1605. 08803. pdf):
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-23.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-24.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-25.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-26.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-27.jpg)
How normalizing flows work? •
![How normalizing flows work How normalizing flows work? •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-28.jpg)
How normalizing flows work? •
![Paper experiments Train the model Glow on one data set in distribution afterwards Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-29.jpg)
Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards determine likelihoods for the training data (in distribution) and another data set that was not used in training (out of distribution)
![Paper experiments Train the model Glow on one data set in distribution afterwards Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-30.jpg)
Paper experiments • Train the model (Glow) on one data set (in distribution), afterwards determine likelihoods for the training data (in distribution) and another data set that was not used in training (out of distribution) • Data set/distribution pairs: • • Fashion. MNIST vs. MNIST CIFAR-10 vs. SVHN Celeb. A vs. SVHN Image. Net vs. CIFAR-10/CIFAR-100/SVHN
![Paper findings Fashion MNIST vs MNIST Paper findings • Fashion. MNIST vs. MNIST](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-31.jpg)
Paper findings • Fashion. MNIST vs. MNIST
![Paper findings Fashion MNIST vs MNIST Paper findings • Fashion. MNIST vs. MNIST](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-32.jpg)
Paper findings • Fashion. MNIST vs. MNIST
![Paper findings CIFAR10 vs SVHN Paper findings • CIFAR-10 vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-33.jpg)
Paper findings • CIFAR-10 vs. SVHN
![Paper findings CIFAR10 vs SVHN Paper findings • CIFAR-10 vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-34.jpg)
Paper findings • CIFAR-10 vs. SVHN
![Paper findings Celeb A vs SVHN Paper findings • Celeb. A vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-35.jpg)
Paper findings • Celeb. A vs. SVHN
![Paper findings Celeb A vs SVHN Paper findings • Celeb. A vs. SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-36.jpg)
Paper findings • Celeb. A vs. SVHN
![Paper findings Image Net vs CIFAR10CIFAR100SVHN Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-37.jpg)
Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN
![Paper findings Image Net vs CIFAR10CIFAR100SVHN Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-38.jpg)
Paper findings • Image. Net vs. CIFAR-10/CIFAR-100/SVHN
![Paper findings Other model types Paper findings • Other model types](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-39.jpg)
Paper findings • Other model types
![Paper findings The observations presented were the main contributions of the paper grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-40.jpg)
Paper findings • The observations presented were the main contributions of the paper, grain of salt needed with next points
![Paper findings The observations presented were the main contributions of the paper grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-41.jpg)
Paper findings • The observations presented were the main contributions of the paper, grain of salt needed with next points • They try to explain the phenomenon, but raising many questions from the reviewers
![Paper findings The observations presented were the main contributions of the paper grain Paper findings • The observations presented were the main contributions of the paper, grain](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-42.jpg)
Paper findings • The observations presented were the main contributions of the paper, grain of salt needed with next points • They try to explain the phenomenon, but raising many questions from the reviewers • Change of variable formula* term analysis:
![Paper findings Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-43.jpg)
Paper findings •
![Paper findings Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-44.jpg)
Paper findings •
![Paper findings Paper findings •](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-45.jpg)
Paper findings •
![Paper findings Then hypothesize that reducing the variance of the data artificially will Paper findings • Then hypothesize that reducing the variance of the data artificially will](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-46.jpg)
Paper findings • Then hypothesize that reducing the variance of the data artificially will increase the likelihood
![Conclusions Cause to pause when using generative models in anomaly detection Second Conclusions • Cause to pause when using generative models in anomaly detection • Second](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-47.jpg)
Conclusions • Cause to pause when using generative models in anomaly detection • Second order analysis provided (only applicable to a certain type of flow + many assumptions) • The author’s urge further study on the subject
![Discussion How validapplicable is their analysis How come samples do not look Discussion • How valid/applicable is their analysis? • How come samples do not look](https://slidetodoc.com/presentation_image_h2/61dd7c14baa5d6ef7ec9e11de7a1ea35/image-48.jpg)
Discussion • How valid/applicable is their analysis? • How come samples do not look like the OOD images if they have higher likelihood?
Normalizing flow
Zhiting hu
Generative vs discriminative
A note on the evaluation of generative models
Taxonomy of generative models
What does sanctioned countries mean
Fake news about nutrition
Fake news
Fake news
Fake news
Craig finlay
Joan naturale
Reflekterende artikel eksempel
Dr michael gans
Goodfellow gan
M*ngulshagai gans*kh
Instabilité
M*ngulshagai gans*kh
Brigitte gans
Applications of gans
Gans loss function
Deep forest: towards an alternative to deep neural networks
深哉深哉耶穌的愛
Deep asleep deep asleep it lies
No news _____ good news.
What is hard news
Probability and counting rules examples with solutions
Deploying deep learning models with docker and kubernetes
Semi modals
Unsupervised image to image translation
Quantum generative adversarial learning
Vb mapp definition
Ontologisk realisme
Generative adversarial networks
Generative design grasshopper
Hudson safety culture
Bayes intranet
Structural linguistic and behavioral psychology
Generative recursion
Generative grammar examples
Generative meditation
Generative lymphoid organs
From structuralism to transformational generative grammar
Lda generative model
Generative adversarial network
Deep and surface structure examples
Nlp generative model
Generative thinking boards
Generative grammar